CN104331978A - Recognition device and method for fold of paper currency - Google Patents

Recognition device and method for fold of paper currency Download PDF

Info

Publication number
CN104331978A
CN104331978A CN201410665409.5A CN201410665409A CN104331978A CN 104331978 A CN104331978 A CN 104331978A CN 201410665409 A CN201410665409 A CN 201410665409A CN 104331978 A CN104331978 A CN 104331978A
Authority
CN
China
Prior art keywords
bank note
fold
grating fringe
value
centerdot
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201410665409.5A
Other languages
Chinese (zh)
Other versions
CN104331978B (en
Inventor
罗攀峰
梁添才
金晓峰
张永
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
GRG Banking Equipment Co Ltd
Guangdian Yuntong Financial Electronic Co Ltd
Original Assignee
Guangdian Yuntong Financial Electronic Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdian Yuntong Financial Electronic Co Ltd filed Critical Guangdian Yuntong Financial Electronic Co Ltd
Priority to CN201410665409.5A priority Critical patent/CN104331978B/en
Publication of CN104331978A publication Critical patent/CN104331978A/en
Priority to PCT/CN2015/087623 priority patent/WO2016078455A1/en
Priority to EP15861703.5A priority patent/EP3223249B1/en
Priority to US15/526,103 priority patent/US10388099B2/en
Priority to RU2017120190A priority patent/RU2645591C1/en
Application granted granted Critical
Publication of CN104331978B publication Critical patent/CN104331978B/en
Priority to ZA2017/03436A priority patent/ZA201703436B/en
Priority to HK18102574.0A priority patent/HK1243215A1/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D7/00—Testing specially adapted to determine the identity or genuineness of valuable papers or for segregating those which are unacceptable, e.g. banknotes that are alien to a currency
    • G07D7/181—Testing mechanical properties or condition, e.g. wear or tear
    • G07D7/183—Detecting folds or doubles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/50—Depth or shape recovery
    • G06T7/521—Depth or shape recovery from laser ranging, e.g. using interferometry; from the projection of structured light
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/50—Depth or shape recovery
    • G06T7/55—Depth or shape recovery from multiple images
    • G06T7/579—Depth or shape recovery from multiple images from motion
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D7/00—Testing specially adapted to determine the identity or genuineness of valuable papers or for segregating those which are unacceptable, e.g. banknotes that are alien to a currency
    • G07D7/06—Testing specially adapted to determine the identity or genuineness of valuable papers or for segregating those which are unacceptable, e.g. banknotes that are alien to a currency using wave or particle radiation
    • G07D7/12—Visible light, infrared or ultraviolet radiation
    • G07D7/121—Apparatus characterised by sensor details
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D7/00—Testing specially adapted to determine the identity or genuineness of valuable papers or for segregating those which are unacceptable, e.g. banknotes that are alien to a currency
    • G07D7/20—Testing patterns thereon
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10141—Special mode during image acquisition
    • G06T2207/10152—Varying illumination
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20076—Probabilistic image processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]

Abstract

The invention relates to a recognition device and a recognition method for the fold of a paper currency. The recognition device for the fold of the paper currency comprises a laser light source, a rectangular raster, an area array photoelectric sensor, an imaging lens assembly and a signal processing module, wherein the laser light source is used for giving out laser light rays, the rectangular raster is located above the laser light source and is used for modulating the laser light rays into stripes which are changed to be bright or dark in accordance with the regular rules on the surface of the paper currency, and the area array photoelectric sensor is used for acquiring images; the imaging lens assembly is located above the area array photoelectric sensor and is used for gathering an image of the surface of the paper currency on the area array photoelectric sensor, and the signal processing module is electrically connected with the area array photoelectric sensor and is used for processing image signals. According to the recognition method for the fold of the paper currency realized by the device is used for converting the inspection whether any fold exists into the inspection whether the feature images of the raster stripes are bent, therefore the influence on the fold of the paper currency caused by a color shift phenomenon due to the change of a grey level of the image acquired of the paper currency at different temperatures can be effectively prevented, and the recognition on the fold of the paper currency can be improved.

Description

The recognition device of bank note fold and method
Technical field
The present invention relates to a kind of non-contact optical detection technique, particularly a kind of recognition device of bank note fold and method.
Background technology
China is a populous country, and the circulation of cash is very huge.Statistics according to People's Bank of China: end the first quarter in 2014, Chinese money in circulation total amount reaches 6.7 trillion yuan Renminbi.Bank note proportion damaged in circulation is large, and being operated in Ge Jia bank to the integral point of bank note, fastidious, point version etc. is all suitable stubborn problem.In order to improve the tidiness of circulation Renminbi, safeguard Renminbi prestige, the People's Bank has formulated " should not circulate the fastidious standard of Renminbi ", wherein there is obviously bank note that is wrinkling or distortion and belongs to one of Renminbi situation that should not circulate.Thus need automatically to identify whether bank note exists buckling problems, and whether the fold degree of this bank note affects it and normally circulates.
The recognition methods to bank note fold of current existence is direct often to be processed bank note collection image itself, be generally judge whether bank note exists fold information by the gray-value variation of bank note collection image neighborhood pixels, the threshold value adopting empirical value or the analysis of a certain amount of bank note sample statistics to determine afterwards is to judge whether fold degree affects the circulation of bank note.Because bank note harvester contact-type image sensor exists certain temperature characterisitic, the banknote image grey scale pixel value gathered in different temperatures situation has certain difference, to bank note be caused to gather image appearance colour cast phenomenon in various degree time serious, thus affect the identification of bank note buckling problems.There is certain uncertainty in single statistical model simultaneously, can not meet the requirement that bank note buckling problems accurately identifies.
Summary of the invention
Certain temperature characterisitic is there is in order to solve contact-type image sensor in prior art, have a strong impact on the identification problem of bank note buckling problems, the invention provides a kind of recognition device of bank note fold, adopt new image acquisition mode, improve the discrimination of bank note fold.
The recognition device of this bank note fold, comprise the transparent protective body that a protective device internal components is clean, wherein, the recognition device of this bank note fold also comprises a LASER Light Source, for sending laser beam; One rectangular raster, it is positioned at the top of this LASER Light Source, for this laser beam is modulated into striped according to the bright dark change of certain rule in note surface; One face array photoelectric sensor, for gathering image; One imaging lens group, is positioned at the top of this face array photoelectric sensor, for note surface picture is converged to this face array photoelectric sensor; And a signal processing module, itself and this face array photoelectric sensor electrically connects, and processes for the picture signal gathered this face array photoelectric sensor.
Preferably, this rectangular raster is rotated by power drive.
The present invention also provides the recognition methods of bank note fold.
The recognition methods of this bank note fold, it comprises: step one, receives the detected bank note of Zou Chao mechanism conveying, and the laser beam that LASER Light Source produces is by rectangular raster and protective, and on negotiable bank note, projection has the grating fringe of the bright dark change of certain rule;
Step 2, face array photoelectric sensor obtains the detected banknote image of this grating fringe additional by imaging lens group, isolates grating fringe characteristic image from the detected banknote image collecting additional grating fringe;
Step 3, rotate this rectangular raster, change LASER Light Source is projected in grating fringe on this detected bank note direction by this rectangular raster, repeated execution of steps two, be separated the grating fringe characteristic image of n angle in the detected banknote image obtaining additional grating fringe, wherein n be more than or equal to 2 integer;
Step 4, message processing module, according to the grating fringe characteristic image of n angle of detected bank note, calculates detected bank note at the graded aggregate-value P of n angle along the grating fringe characteristic image in grating fringe direction; And
Step 5, the graded aggregate-value P of this detected bank note and the decision threshold T pre-set are compared, provide the result whether detected bank note exists fold, concrete comparison procedure is as follows: if P≤T, then bank note does not exist fold, allows circulation; If P > is T, then there is fold in bank note, do not allow circulation.
Preferably, this decision threshold T obtains as follows: step one: the negotiable bank note receiving the conveying of Zou Chao mechanism, the laser beam that LASER Light Source produces is by rectangular raster and protective, and on negotiable bank note, projection has the grating fringe of the bright dark change of certain rule;
Step 2: face array photoelectric sensor obtains the negotiable banknote image of this grating fringe additional by imaging lens group, isolates grating fringe characteristic image from the negotiable banknote image collecting additional grating fringe;
Step 3: rotate this rectangular raster, change LASER Light Source is projected in grating fringe on this negotiable bank note direction by this rectangular raster, repeated execution of steps two, be separated the grating fringe characteristic image of n angle in the negotiable banknote image obtaining additional grating fringe, wherein n be more than or equal to 2 integer;
Step 4: message processing module, according to the grating fringe characteristic image of the n of a negotiable bank note angle, calculates negotiable bank note at the graded aggregate-value S of n angle along the grating fringe characteristic image in grating fringe direction;
Step 5: message processing module opens negotiable bank note sample according to U, performs step one respectively to step 4, obtains each negotiable bank note with the graded aggregate-value Sa of the grating fringe image along grating fringe direction, wherein a=1.2 ... U;
Step 6, using graded aggregate-value Sa as feature, message processing module adopts Statistic analysis models and/or artificial neural network learning training method to obtain the decision threshold T that whether there is fold.
Preferably, negotiable bank note refer to meet People's Bank of China's regulation can normal paper money in circulation sample; The grating fringe with the bright dark conversion of certain rule refers to that this laser beam modulates through this rectangular raster the striped formed.
Preferably, this n be more than or equal to 8 integer, the anglec of rotation of this rectangular raster is angle uniform subdivision within the scope of 360 degree being become n-1 part, as for 8 angles, these 8 angles refer to that the anglec of rotation of rectangular raster is respectively 0 degree, 45 degree, 90 degree, 135 degree, 180 degree, 225 degree, 270 degree and 315 degree, i.e. the anglec of rotation of the rectangular raster angle that will evenly be divided into 7 parts within the scope of 360 degree.
Preferably, as follows along the computing formula of graded aggregate-value P and S of the grating fringe characteristic image in grating fringe direction n angle:
P = Σ i = 1 n Σ j = 1 M Σ K = 1 N [ P ( i , j , K + 1 ) - P ( i , j , K ) ] ;
S = Σ i = 1 n Σ j = 1 M Σ K = 1 N [ P ( i , j , K + 1 ) - P ( i , j , K ) ] ;
Wherein, i is the rectangular raster rotation angle number of degrees, n be more than or equal to 2 integer, M is the quantity of the bright dark change of grating fringe characteristic image, N is the quantity along pixel on the single striped on stripe direction, P (i, j, k) is the pixel value of K pixel on grating fringe i-th angle acquisition image jth striped.
Preferably, described Statistic analysis models is that in analysis of range, variance analysis and interval estimation weighted analysis, at least one is formed.
Preferably, the specific formula for calculation of analysis of range value is as follows:
r = max ( S a ) - min ( S a ) 2 , a = 1.2 · · · · · · U ; Wherein analysis of range value r corresponds to the fold maximal value of negotiable bank note sample and the half of minimum value, and U is the quantity of bank note sample.
Preferably, the concrete formula of variance analysis is: wherein variance analysis value f corresponds to the fold degree of fluctuation of negotiable bank note sample, and U is the quantity of bank note sample, be the mean value of the graded aggregate-value Sa of negotiable bank note sample, concrete formula is: S ‾ = Σ a = 1 U S ( a ) U , a = 1.2 · · · · · · U .
Preferably, the concrete formula of interval estimation weighted analysis is: this interval estimation weighted value q corresponds to the fold parameter estimation of negotiable bank note sample, and wherein, U is the quantity of bank note sample, Z 1-bbe confidence factor, 1-b is confidence level, and d is scale parameter.
Preferably, message processing module adopts Statistic analysis models and/or artificial neural network learning training method to obtain the decision threshold T that whether there is fold, concrete steps are, message processing module adopts Artificial Neural Network to the statistical value r of analysis of range, the statistical value f of variance analysis, the value q of interval estimation weighted analysis carries out learning training, utilize the error of the directly front conducting shell of the estimation of error output layer after exporting, again by the error of the more front one deck of this estimation of error, so in layer anti-pass, obtain the estimation of error of every other each layer, artificial neural network constantly changes the connection weights of network under the stimulation of extraneous input amendment, to make the output of network constantly close to the output expected, finally obtain analysis of range weighted value ω 1, variance analysis weighted value ω 2with interval estimation weighted value ω 3, thus obtain striped on raster image and whether there is bending graded accumulated value threshold value T, concrete formula is as follows:
T = S ‾ + ω 1 r + ω 2 f + ω 3 q .
The recognition methods of bank note fold provided by the present invention has effective as follows:
The recognition methods of bank note fold provided by the invention, whether the grating fringe image that the detection whether bank note being existed fold is transformed into projection grating exists bending detection, effectively prevent bank note in different temperatures situation and gather colour cast phenomenon that the gray-value variation of image own causes to the impact of bank note fold identification, improve bank note fold discrimination.The method is according to negotiable bank note sample, using the graded aggregate-value of the grating fringe image along stripe direction as feature, adopt the method for multiple Statistical parameter analysis and learning training to determine threshold value that whether buckle condition affects fiduciary circulation, achieves the accurate identification of bank note fold.Compared with adopting the method for empirical value or single Statistical parameter analysis definite threshold before, the method for the threshold value that the present invention determines is more true and reliable, has stronger universality.
Accompanying drawing explanation
Fig. 1 is the recognition device structural representation of bank note fold;
Fig. 2 is the rectangular raster striped schematic diagram of projection;
Fig. 3 is the grating fringe image wrapping schematic diagram that bank note fold causes.
Embodiment
For setting forth bank note fold provided by the present invention recognition methods and device further, present embodiment is described in detail in conjunction with diagram.
Present embodiments provide a kind of recognition methods to bank note fold; as shown in Figure 1, the recognition device of this bank note fold comprises transparent protective body 1, rectangular raster 3, LASER Light Source 4, imaging lens group 5, face array photoelectric sensor 6 and message processing module 7 to the recognition device of the implement device bank note fold of the method.LASER Light Source 4 is for sending laser beam; Rectangular raster 3 is positioned at the top of LASER Light Source 4, for laser beam 4 is modulated into striped according to the bright dark change of certain rule in note surface; Face array photoelectric sensor 6 is for the collection of image; Imaging lens group 5 is positioned at the top of face array photoelectric sensor 6, for note surface picture is converged to face array photoelectric sensor 6; Signal processing module 7 and face array photoelectric sensor 6 electrically connect, and process for the picture signal gathered this face array photoelectric sensor 6.Wherein protective 1 is for keeping the clean of internal components, be clear glass in this protective 1 the present embodiment, it is to be noted the sheet material can also with the transparent material formation of some strength, as transparent biochip etc.
According to the description corresponding with the recognition device of above-mentioned bank note fold, the recognition methods of the bank note fold that the present embodiment provides comprises:
Step one, receives the detected bank note of Zou Chao mechanism conveying, and the laser beam that LASER Light Source 4 produces is by rectangular raster 3 and protective 1, and on detected bank note, projection has the grating fringe of the bright dark change of certain rule; It is such as the rectangular raster striped schematic diagram of projection shown in Fig. 2; Fig. 3 is the grating fringe image wrapping schematic diagram that bank note fold causes;
Step 2, face array photoelectric sensor 6 obtains the detected banknote image of this grating fringe additional by imaging lens group 5, isolates grating fringe characteristic image from the detected banknote image collecting additional grating fringe;
Step 3, rotate this rectangular raster 3, change LASER Light Source is projected in grating fringe on this detected bank note direction by this rectangular raster, repeated execution of steps two, be separated the grating fringe characteristic image of n angle in the detected banknote image obtaining additional grating fringe, wherein n be more than or equal to 2 integer;
When the direction of bank note fold and the grating fringe direction of projection consistent time, the grating fringe collected can not reflect the buckling phenomena in this direction, thus in order to effectively detect the omnidirectional fold information of bank note, need to gather the projection grating stripe pattern being more than or equal to 2 angles, preferably this n be more than or equal to 8 integer, the anglec of rotation of this rectangular raster is angle uniform subdivision within the scope of 360 degree being become n-1 part, as for 8 angles, refer to the angle being separated the single pixel center of raster image and the line correspondences at adjacent pixel center place obtained, these 8 angles refer to that the anglec of rotation of rectangular raster is respectively 0 degree, 45 degree, 90 degree, 135 degree, 180 degree, 225 degree, 270 degree and 315 degree, the i.e. anglec of rotation of the rectangular raster angle that will evenly be divided into 7 parts within the scope of 360 degree.When n is for being greater than 8 angles, then further uniform subdivision on 0 degree, 45 degree, 90 degree, 135 degree, 180 degree, 225 degree, 270 degree, 315 degree bases, namely within the scope of 360 degree, uniform subdivision becomes the angle of n-1 part, for 12 angles, the anglec of rotation of rectangular raster can be arranged respectively to 0 degree, 30 degree, 60 degree, 90 degree, 120 degree, 150 degree, 180 degree, 210 degree, 240 degree, 270 degree, 300 degree, 330 degree, becomes the angle of 11 parts by uniform subdivision within the scope of 360 degree.
Step 4, message processing module 7, according to the grating fringe characteristic image of the n of a detected bank note angle, calculates detected bank note at the graded aggregate-value P of n angle along the grating fringe characteristic image in grating fringe direction, P = Σ i = 1 n Σ j = 1 M Σ K = 1 N [ P ( i , j , K + 1 ) - P ( i , j , K ) ] , Wherein, i is the rectangular raster rotation angle number of degrees, n be more than or equal to 2 integer, M is the quantity of the bright dark change of grating fringe characteristic image, N is the quantity along pixel on the single striped on stripe direction, P (i, j, k) is the pixel value of K pixel on grating fringe i-th angle acquisition image jth striped.
Step 5, the graded aggregate-value P of this detected bank note and the decision threshold T pre-set are compared, provide the result whether detected bank note exists fold, concrete comparison procedure is as follows: if P≤T, then bank note does not exist fold, allows circulation; If P > is T, then there is fold in bank note, do not allow circulation.
In order to realize above-mentioned recognition methods, need to determine decision threshold T in advance, the determining step of this decision threshold T is as follows:
Step one: the negotiable bank note receiving the conveying of Zou Chao mechanism, the laser beam that LASER Light Source 4 produces is by rectangular raster 3 and protective 1, and on negotiable bank note 2, projection has the grating fringe of the bright dark change of certain rule; Wherein negotiable bank note refer to meet the People's Bank's regulation can normal paper money in circulation sample, the grating fringe with the bright dark conversion of certain rule refers to that laser beam modulates the striped formed through rectangular raster, such as, be the rectangular raster striped schematic diagram of projection shown in Fig. 2; Fig. 3 is the grating fringe image wrapping schematic diagram that bank note fold causes.
Step 2: face array photoelectric sensor 6 obtains the negotiable banknote image of additional grating fringe by imaging lens group 5, isolates grating fringe characteristic image from the negotiable banknote image collecting additional grating fringe.
Step 3: rotate rectangular raster 3, change LASER Light Source is projected in grating fringe on this negotiable bank note direction by rectangular raster, repeated execution of steps two, be separated the grating fringe characteristic image of n angle in the negotiable banknote image obtaining additional grating fringe, wherein n be more than or equal to 2 integer;
When the direction of bank note fold and the grating fringe direction of projection consistent time, the grating fringe collected can not reflect the buckling phenomena in this direction, thus in order to effectively detect the omnidirectional fold information of bank note, need to gather the projection grating stripe pattern being more than or equal to 2 angles, preferably this n be more than or equal to 8 integer, the anglec of rotation of this rectangular raster is angle uniform subdivision within the scope of 360 degree being become n-1 part, as for 8 angles, refer to the angle being separated the single pixel center of raster image and the line correspondences at adjacent pixel center place obtained, these 8 angles refer to that the anglec of rotation of rectangular raster is respectively 0 degree, 45 degree, 90 degree, 135 degree, 180 degree, 225 degree, 270 degree and 315 degree, the i.e. anglec of rotation of the rectangular raster angle that will evenly be divided into 7 parts within the scope of 360 degree.When n is for being greater than 8 angles, then further uniform subdivision on 0 degree, 45 degree, 90 degree, 135 degree, 180 degree, 225 degree, 270 degree, 315 degree bases, namely within the scope of 360 degree, uniform subdivision becomes the angle of n-1 part, for 12 angles, the anglec of rotation of rectangular raster can be arranged respectively to 0 degree, 30 degree, 60 degree, 90 degree, 120 degree, 150 degree, 180 degree, 210 degree, 240 degree, 270 degree, 300 degree, 330 degree, becomes the angle of 11 parts by uniform subdivision within the scope of 360 degree.
Step 4: message processing module 7 is according to the grating fringe characteristic image of the n of a negotiable bank note angle, and calculate negotiable bank note at the graded aggregate-value S of n angle along the grating fringe characteristic image in grating fringe direction, concrete mathematical formulae is as follows:
S = Σ i = 1 n Σ j = 1 M Σ K = 1 N [ P ( i , j , K + 1 ) - P ( i , j , K ) ] ;
Wherein, i is the rectangular raster anglec of rotation, n be more than or equal to 2 integer, M is the quantity of the bright dark change of grating fringe image projecting to note surface, N is the quantity along pixel on the single striped on stripe direction, P (i, j, k) is the pixel value of K pixel on grating fringe i-th angle acquisition image jth striped.
Step 5: message processing module 7 opens negotiable bank note sample according to U, perform step one respectively to step 4, obtain each negotiable bank note with the graded aggregate-value Sa of the grating fringe characteristic image along grating fringe direction, wherein a is natural number, a=1.2 ... U, and U>=200, this U value is larger better for recognition effect.
Step 6: using graded aggregate-value Sa as feature, adopt analysis of range, variance analysis and the modeling statistics analysis of interval estimation weighting composition and the method determination bank note of neural network learning training whether to there is the decision threshold T of fold, detailed process is as follows:
Message processing module 7 obtains analysis of range value r, variance analysis value f, interval estimation weighted value q tri-statistical natures of the graded aggregate-value Sa of negotiable bank note sample, and concrete formula is as follows:
Analysis of range: r = max ( S a ) - min ( S a ) 2 , a = 1.2 · · · · · · U ; Wherein correspond to can the fold maximal value of circulating paper money sample and the half of minimum value for analysis of range value r, and U is the quantity of bank note sample.
Variance analysis: variance analysis corresponds to the fold degree of fluctuation of negotiable bank note sample to f, and U is the quantity of bank note sample, wherein be can the mean value of graded aggregate-value of circulating paper money sample, concrete formula be as follows:
S ‾ = Σ a = 1 U S ( a ) U .
Interval estimation weighting: this interval estimation weighted value q corresponds to the fold parameter estimation of negotiable bank note sample, and wherein, U is the quantity of bank note sample, Z 1-bbe confidence factor, 1-b is confidence level, and d is scale parameter.
Compared with adopting the method for empirical value or single Statistical parameter analysis determination decision threshold T, bank note fold recognition methods that the present embodiment provides adopts analysis of range, variance analysis and interval estimation weighted analysis three parameters to characterize can the buckle condition of circulating paper money, more comprehensively can reflect the buckle condition of bank note, judged result is more true and reliable, has stronger universality.
Message processing module 7 adopts Artificial Neural Network to carry out learning training to statistical value r, the statistical value f of variance analysis of analysis of range, the value q of interval estimation weighted analysis, ultimate principle utilizes the error after exporting to estimate the error of the directly front conducting shell of output layer, again by the error of the more front one deck of this estimation of error, anti-pass so is in layer gone down, and just obtains the estimation of error of every other each layer.Neural network constantly changes the connection weights of network under the stimulation of extraneous input amendment, to make the output of network constantly close to the output expected, finally obtains analysis of range weighted value ω 1, variance analysis weighted value ω 2with interval estimation weighted value ω 3, thus striped on raster image can be obtained whether there is bending graded accumulated value threshold value T, concrete formula is as follows:
T = S ‾ + ω 1 r + ω 2 f + ω 3 q .
Whether the grating fringe characteristic image that the detection whether bank note is existed fold by the present invention is transformed into projection grating exists bending detection, effectively prevent bank note in different temperatures situation and gathers colour cast phenomenon that the gray-value variation of image own causes to the impact of bank note fold identification.The present invention is according to can paper money in circulation sample, using the graded aggregate-value of the grating fringe image along stripe direction as feature, adopt the method for multiple Statistical parameter analysis and learning training to determine decision threshold that whether buckle condition affects fiduciary circulation, achieves effective identification of bank note fold.Compared with adopting the method for empirical value or single Statistical parameter analysis determination decision threshold before, the method for the threshold value that the present invention determines is more true and reliable, has stronger universality.
Below be only the preferred embodiment of the present invention, it should be pointed out that above-mentioned preferred implementation should not be considered as limitation of the present invention, protection scope of the present invention should be as the criterion with claim limited range.For those skilled in the art, without departing from the spirit and scope of the present invention, can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (12)

1. a recognition device for bank note fold, comprise the transparent protective body that a protective device internal components is clean, it is characterized in that, the recognition device of this bank note fold also comprises:
One LASER Light Source, for sending laser beam;
One rectangular raster, it is positioned at the top of this LASER Light Source, for this laser beam is modulated into striped according to the bright dark change of certain rule in note surface;
One face array photoelectric sensor, for gathering image;
One imaging lens group, is positioned at the top of this face array photoelectric sensor, for note surface picture is converged to this face array photoelectric sensor; And
One signal processing module, itself and this face array photoelectric sensor electrically connects, and processes for the picture signal gathered this face array photoelectric sensor.
2. the recognition device of bank note fold as claimed in claim 1, it is characterized in that, this rectangular raster is rotated by power drive.
3. a recognition methods for bank note fold, it comprises:
Step one, receives the detected bank note of Zou Chao mechanism conveying, and the laser beam that LASER Light Source produces is by rectangular raster and protective, and on detected bank note, projection has the grating fringe of the bright dark change of certain rule;
Step 2, face array photoelectric sensor obtains the detected banknote image of this grating fringe additional by imaging lens group, isolates grating fringe characteristic image from the detected banknote image collecting additional grating fringe;
Step 3, rotate this rectangular raster, change LASER Light Source is projected in grating fringe on this detected bank note direction by this rectangular raster, repeated execution of steps two, be separated the grating fringe characteristic image of n angle in the detected banknote image obtaining additional grating fringe, wherein n be more than or equal to 2 integer;
Step 4, message processing module, according to the grating fringe characteristic image of n angle of detected bank note, calculates detected bank note at the graded aggregate-value P of n angle along the grating fringe characteristic image in grating fringe direction; And
Step 5, compares the graded aggregate-value P of this detected bank note and the decision threshold T pre-set, provides the result whether detected bank note exists fold.
4. the recognition methods of bank note fold as claimed in claim 3, it is characterized in that, this decision threshold T obtains as follows:
Step one: the negotiable bank note receiving the conveying of Zou Chao mechanism, the laser beam that LASER Light Source produces is by rectangular raster and protective, and on negotiable bank note, projection has the grating fringe of the bright dark change of certain rule;
Step 2: face array photoelectric sensor obtains the negotiable banknote image of this grating fringe additional by imaging lens group, isolates grating fringe characteristic image from the negotiable banknote image collecting additional grating fringe;
Step 3: rotate this rectangular raster, change LASER Light Source is projected in grating fringe on this negotiable bank note direction by this rectangular raster, repeated execution of steps two, be separated the grating fringe characteristic image of n angle in the negotiable banknote image obtaining additional grating fringe, wherein n be more than or equal to 2 integer;
Step 4: message processing module, according to the grating fringe characteristic image of the n of a negotiable bank note angle, calculates negotiable bank note at the graded aggregate-value S of n angle along the grating fringe characteristic image in grating fringe direction;
Step 5: message processing module opens negotiable bank note sample according to U, performs step one respectively to step 4, obtains each negotiable bank note with the graded aggregate-value Sa of the grating fringe image along grating fringe direction, wherein a=1.2 ... U;
Step 6, using graded aggregate-value Sa as feature, message processing module adopts Statistic analysis models and/or artificial neural network learning training method to obtain the decision threshold T that whether there is fold.
5. the recognition methods of bank note fold as claimed in claim 4, is characterized in that, negotiable bank note refer to meet People's Bank of China's regulation can normal paper money in circulation sample; The grating fringe with the bright dark conversion of certain rule refers to that this laser beam modulates through this rectangular raster the striped formed.
6. the recognition methods of the bank note fold as described in claim 3 or 4, is characterized in that, this n be more than or equal to 8 integer, the anglec of rotation of this rectangular raster is angle uniform subdivision within the scope of 360 degree being become n-1 part.
7. the recognition methods of the bank note fold as described in claim 3 or 4, is characterized in that, as follows along the computing formula of graded aggregate-value P or S of the grating fringe characteristic image in grating fringe direction n angle:
PorS = Σ i = 1 n Σ j = 1 M Σ K = 1 N [ P ( i , j , K + 1 ) - P ( i , j , K ) ]
Wherein, i is the rectangular raster rotation angle number of degrees, n be more than or equal to 2 integer, M is the quantity of the bright dark change of grating fringe characteristic image, N is the quantity along pixel on the single striped on stripe direction, P (i, j, k) is the pixel value of K pixel on grating fringe i-th angle acquisition image jth striped.
8. the recognition methods of bank note fold as claimed in claim 4, is characterized in that, described Statistic analysis models is that in analysis of range, variance analysis and interval estimation weighted analysis, at least one is formed.
9. the recognition methods of bank note fold as claimed in claim 8, it is characterized in that, the concrete formula of described analysis of range is as follows:
r = max ( S a ) - min ( S a ) 2 , a = 1.2 · · · · · · U ;
Wherein analysis of range value r corresponds to the fold maximal value of negotiable bank note sample and the half of minimum value, and U is the quantity of bank note sample.
10. the recognition methods of bank note fold as claimed in claim 8, it is characterized in that, the concrete formula of described variance analysis is as follows:
f = Σ a = 1 U [ S a - S ‾ ] , a = 1.2 · · · · · · U
Wherein variance analysis value f corresponds to the fold degree of fluctuation of negotiable bank note sample, and U is the quantity of bank note sample, be the mean value of the graded aggregate-value Sa of negotiable bank note sample, concrete formula is as follows:
S ‾ = Σ a = 1 U S ( a ) U , a = 1.2 · · · · · · U .
The recognition methods of 11. bank note folds as claimed in claim 8, it is characterized in that, the concrete formula of described interval estimation weighted analysis is as follows:
q = Z 1 - b d U ,
This interval estimation weighted value q corresponds to the fold estimates of parameters of negotiable bank note sample, and wherein, U is the quantity of bank note sample, Z 1-bbe confidence factor, 1-b is confidence level, and d is scale parameter.
The recognition methods of 12. bank note folds as claimed in claim 8, it is characterized in that, step 6 is specially, message processing module adopts Artificial Neural Network respectively to the statistical value r of analysis of range, the statistical value f of variance analysis, the value q of interval estimation weighted analysis carries out learning training, utilize the error of the directly front conducting shell of the estimation of error output layer after exporting, again by the error of the more front one deck of this estimation of error, so in layer anti-pass, obtain the estimation of error of all each layers, artificial neural network constantly changes the connection weights of network under the stimulation of extraneous input amendment, to make the output of network constantly close to the output expected, finally obtain analysis of range weighted value ω 1, variance analysis weighted value ω 2with interval estimation weighted value ω 3, thus obtain striped on raster image and whether there is bending graded accumulated value threshold value T, concrete formula is as follows:
T = S ‾ + ω 1 r + ω 2 f + ω 3 q .
CN201410665409.5A 2014-11-19 2014-11-19 Recognition device and method for fold of paper currency Active CN104331978B (en)

Priority Applications (7)

Application Number Priority Date Filing Date Title
CN201410665409.5A CN104331978B (en) 2014-11-19 2014-11-19 Recognition device and method for fold of paper currency
PCT/CN2015/087623 WO2016078455A1 (en) 2014-11-19 2015-08-20 Paper currency fold recognition apparatus and method
EP15861703.5A EP3223249B1 (en) 2014-11-19 2015-08-20 Paper currency fold recognition apparatus and method
US15/526,103 US10388099B2 (en) 2014-11-19 2015-08-20 Paper currency fold recognition apparatus and method
RU2017120190A RU2645591C1 (en) 2014-11-19 2015-08-20 Device and method of recognizing bend paper of money
ZA2017/03436A ZA201703436B (en) 2014-11-19 2017-05-18 Paper currency fold recognition apparatus and method
HK18102574.0A HK1243215A1 (en) 2014-11-19 2018-02-22 Paper currency fold recognition apparatus and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410665409.5A CN104331978B (en) 2014-11-19 2014-11-19 Recognition device and method for fold of paper currency

Publications (2)

Publication Number Publication Date
CN104331978A true CN104331978A (en) 2015-02-04
CN104331978B CN104331978B (en) 2017-02-01

Family

ID=52406697

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410665409.5A Active CN104331978B (en) 2014-11-19 2014-11-19 Recognition device and method for fold of paper currency

Country Status (7)

Country Link
US (1) US10388099B2 (en)
EP (1) EP3223249B1 (en)
CN (1) CN104331978B (en)
HK (1) HK1243215A1 (en)
RU (1) RU2645591C1 (en)
WO (1) WO2016078455A1 (en)
ZA (1) ZA201703436B (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105096446A (en) * 2015-07-20 2015-11-25 深圳怡化电脑股份有限公司 Method and system for identifying and authenticating optically variable ink on banknotes with folding corners
WO2016078455A1 (en) * 2014-11-19 2016-05-26 广州广电运通金融电子股份有限公司 Paper currency fold recognition apparatus and method
CN106355739A (en) * 2016-08-18 2017-01-25 深圳怡化电脑股份有限公司 Method and device for detecting new or old paper money
CN107393131A (en) * 2017-07-25 2017-11-24 深圳怡化电脑股份有限公司 Note detection device and cash automated trading device
CN108074329A (en) * 2018-01-02 2018-05-25 深圳怡化电脑股份有限公司 Paper money processing system, method, apparatus, equipment and storage medium
CN108318355A (en) * 2017-12-31 2018-07-24 广德大金机械有限公司 A kind of leather and fur products folding quality intelligent test method based on comparative analysis
CN110849911A (en) * 2019-11-25 2020-02-28 厦门大学 Glass defect image acquisition device, glass defect detection equipment and detection method
CN111660692A (en) * 2020-04-28 2020-09-15 深圳大学 Financial document intelligent processing system and device based on multi-wavelength optical fold identification
CN113899755A (en) * 2021-11-17 2022-01-07 武汉华星光电半导体显示技术有限公司 Screen crease degree detection method and visual detection equipment

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10616443B1 (en) * 2019-02-11 2020-04-07 Open Text Sa Ulc On-device artificial intelligence systems and methods for document auto-rotation

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6522399B1 (en) * 1997-02-24 2003-02-18 Qinetiq Limited Signature mark recognition systems
CN1892710A (en) * 2005-07-06 2007-01-10 日立欧姆龙金融系统有限公司 Treatment of paper money with topical folding thereon
CN101901511A (en) * 2009-05-27 2010-12-01 株式会社东芝 Document handling apparatus
CN102142165A (en) * 2010-12-31 2011-08-03 南京理工速必得科技股份有限公司 Device for acquiring paper currency face feature images at high speed
CN102568081A (en) * 2012-01-12 2012-07-11 浙江大学 Image acquisition and processing method and device of paper money discriminator
EP2546808A1 (en) * 2011-07-13 2013-01-16 Glory Ltd. Paper sheet recognition apparatus and paper sheet recognition method

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS575768U (en) * 1980-06-12 1982-01-12
WO2002050783A1 (en) 2000-12-21 2002-06-27 Cambridge Consultants Limited Optical sensor device and method for spectral analysis
DE10159234B4 (en) 2001-12-03 2012-12-13 Giesecke & Devrient Gmbh Device for examining documents
DE10323409A1 (en) * 2003-05-23 2004-12-09 Giesecke & Devrient Gmbh Device for checking banknotes
US7184149B2 (en) 2003-06-18 2007-02-27 Dimensional Photonics International, Inc. Methods and apparatus for reducing error in interferometric imaging measurements
JP4566543B2 (en) * 2003-10-10 2010-10-20 日本金銭機械株式会社 Valuable paper sheet identification device
JP2006284212A (en) 2005-03-31 2006-10-19 Dainippon Screen Mfg Co Ltd Unevenness inspection device and unevenness inspection method
WO2007044570A2 (en) * 2005-10-05 2007-04-19 Cummins-Allison Corp. Currency processing system with fitness detection
CN100504290C (en) 2006-06-20 2009-06-24 东华大学 3D non-contacting type coordinates equipment for measuring fabric or garment material
CN200976166Y (en) 2006-09-26 2007-11-14 天津雪赢世纪专利技术有限公司 Banknote detecting card
CN101451826B (en) 2008-12-17 2010-06-09 中国科学院上海光学精密机械研究所 Object three-dimensional contour outline measuring set and measuring method
WO2012116807A1 (en) 2011-03-03 2012-09-07 Optocraft Gmbh Method and device for topography measurement on the eye
JP5762830B2 (en) * 2011-06-07 2015-08-12 グローリー株式会社 Paper sheet processing apparatus and method
CN103115586A (en) 2013-02-05 2013-05-22 华南理工大学 Micro three-dimensional sensing device based on laser interference fringes
CN104331978B (en) 2014-11-19 2017-02-01 广州广电运通金融电子股份有限公司 Recognition device and method for fold of paper currency

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6522399B1 (en) * 1997-02-24 2003-02-18 Qinetiq Limited Signature mark recognition systems
CN1892710A (en) * 2005-07-06 2007-01-10 日立欧姆龙金融系统有限公司 Treatment of paper money with topical folding thereon
CN101901511A (en) * 2009-05-27 2010-12-01 株式会社东芝 Document handling apparatus
CN102142165A (en) * 2010-12-31 2011-08-03 南京理工速必得科技股份有限公司 Device for acquiring paper currency face feature images at high speed
EP2546808A1 (en) * 2011-07-13 2013-01-16 Glory Ltd. Paper sheet recognition apparatus and paper sheet recognition method
CN102568081A (en) * 2012-01-12 2012-07-11 浙江大学 Image acquisition and processing method and device of paper money discriminator

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016078455A1 (en) * 2014-11-19 2016-05-26 广州广电运通金融电子股份有限公司 Paper currency fold recognition apparatus and method
US10388099B2 (en) 2014-11-19 2019-08-20 Grg Banking Equipment Co., Ltd. Paper currency fold recognition apparatus and method
CN105096446A (en) * 2015-07-20 2015-11-25 深圳怡化电脑股份有限公司 Method and system for identifying and authenticating optically variable ink on banknotes with folding corners
CN106355739A (en) * 2016-08-18 2017-01-25 深圳怡化电脑股份有限公司 Method and device for detecting new or old paper money
CN106355739B (en) * 2016-08-18 2019-03-12 深圳怡化电脑股份有限公司 A kind of method and device that detection bank note is new and old
CN107393131A (en) * 2017-07-25 2017-11-24 深圳怡化电脑股份有限公司 Note detection device and cash automated trading device
CN108318355A (en) * 2017-12-31 2018-07-24 广德大金机械有限公司 A kind of leather and fur products folding quality intelligent test method based on comparative analysis
CN108074329A (en) * 2018-01-02 2018-05-25 深圳怡化电脑股份有限公司 Paper money processing system, method, apparatus, equipment and storage medium
CN110849911A (en) * 2019-11-25 2020-02-28 厦门大学 Glass defect image acquisition device, glass defect detection equipment and detection method
CN111660692A (en) * 2020-04-28 2020-09-15 深圳大学 Financial document intelligent processing system and device based on multi-wavelength optical fold identification
CN111660692B (en) * 2020-04-28 2024-03-01 深圳大学 Financial document intelligent processing system and device based on multi-wavelength optical fold recognition
CN113899755A (en) * 2021-11-17 2022-01-07 武汉华星光电半导体显示技术有限公司 Screen crease degree detection method and visual detection equipment

Also Published As

Publication number Publication date
CN104331978B (en) 2017-02-01
WO2016078455A1 (en) 2016-05-26
EP3223249A4 (en) 2017-11-29
RU2645591C1 (en) 2018-02-21
HK1243215A1 (en) 2018-07-06
EP3223249B1 (en) 2021-12-08
ZA201703436B (en) 2019-07-31
US10388099B2 (en) 2019-08-20
US20170316632A1 (en) 2017-11-02
EP3223249A1 (en) 2017-09-27

Similar Documents

Publication Publication Date Title
CN104331978A (en) Recognition device and method for fold of paper currency
CN109410207A (en) A kind of unmanned plane line walking image transmission line faultlocating method based on NCC feature
CN103679636B (en) Based on point, the fast image splicing method of line double characteristic
CN104112269B (en) A kind of solar battery laser groove parameter detection method and system based on machine vision
CN110263705A (en) Towards two phase of remote sensing technology field high-resolution remote sensing image change detecting method
CN106127204A (en) A kind of multi-direction meter reading Region detection algorithms of full convolutional neural networks
CN101930533B (en) Device and method for performing sky detection in image collecting device
CN110232389A (en) A kind of stereoscopic vision air navigation aid based on green crop feature extraction invariance
CN103281513B (en) Pedestrian recognition method in the supervisory control system of a kind of zero lap territory
CN106680285A (en) Infrared image-assisted method of recognizing contamination condition of insulator by visible light image
CN104408473B (en) Cotton grade sorting technique and device based on learning distance metric
CN103164695A (en) Fruit identification method based on multi-source image information fusion
CN103034838A (en) Special vehicle instrument type identification and calibration method based on image characteristics
CN104240204A (en) Solar silicon wafer and battery piece counting method based on image processing
Yuan et al. Learning to count buildings in diverse aerial scenes
CN105894513B (en) Take the remote sensing image variation detection method and system of imaged object change in time and space into account
CN103679677A (en) Dual-model image decision fusion tracking method based on mutual updating of models
CN108765476A (en) A kind of polarization image method for registering
CN107065037A (en) A kind of Data of Automatic Weather acquisition control system
CN104680151B (en) A kind of panchromatic remote sensing image variation detection method of high-resolution for taking snow covering influence into account
CN111507426A (en) No-reference image quality grading evaluation method and device based on visual fusion characteristics
CN105139375A (en) Satellite image cloud detection method combined with global DEM and stereo vision
CN107194393A (en) A kind of method and device for detecting Provisional Number Plate
CN204039886U (en) A kind of pavement damage crack detection system based on multiple stage camera stereoscopic shooting
CN109115719A (en) A kind of Citrus Huanglongbing pathogen Band fusion rapid detection method based on high light spectrum image-forming technology

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
PE01 Entry into force of the registration of the contract for pledge of patent right
PE01 Entry into force of the registration of the contract for pledge of patent right

Denomination of invention: Device and method for identifying paper currency wrinkles

Effective date of registration: 20230811

Granted publication date: 20170201

Pledgee: Bank of China Limited by Share Ltd. Guangzhou Tianhe branch

Pledgor: GRG BANKING EQUIPMENT Co.,Ltd.

Registration number: Y2023980051844